Markerless Kinematic Data in the Frontal Plane Contributions to Movement Quality in the Single-Leg Squat Test: A Comparison and Decision Tree Approach.

OBJECTIVE

The aim of this study is to compare kinematic data of the frontal trunk, pelvis, knee, and summated angles (trunk plus knee) among categorized grades using the single-leg squat (SLS) test, to classify the SLS grade, and to investigate the association between the SLS grade and the frontal angles using smartphone-based markerless motion capture.

METHODS

Ninety-one participants were categorized into 3 grades (good, reduced, and poor) based on the quality of the SLS test. An automated pose estimation algorithm was employed to assess the frontal joint angles during SLS, which were captured by a single smartphone camera. Analysis of variance and a decision tree model using classification and regression tree analysis were utilized to investigate intergroup differences, classify the SLS grades, and identify associations between the SLS grade and frontal angles, respectively.

RESULTS

In the poor group, each frontal trunk, knee, and summated angle was significantly larger than in the good group. Classification and regression tree analysis showed that frontal knee and summated angles could classify the SLS grades with a 76.9% accuracy. Additionally, the classification and regression tree analysis established cutoff points for each frontal knee (11.34°) and summated angles (28.4°), which could be used in clinical practice to identify individuals who have a reduced or poor grade in the SLS test.

CONCLUSIONS

The quality of SLS was found to be associated with interactions among frontal knee and summated angles. With an automated pose estimation algorithm, a single smartphone computer vision method can be utilized to compare and distinguish the quality of SLS movement for remote clinical and sports assessments.